xen/netfront: don't bug in case of too many frags

Commit 57f230a ("xen/netfront: raise max number of slots in
xennet_get_responses()") raised the max number of allowed slots by one.
This seems to be problematic in some configurations with netback using
a larger MAX_SKB_FRAGS value (e.g. old Linux kernel with MAX_SKB_FRAGS
defined as 18 instead of nowadays 17).

Instead of BUG_ON() in this case just fall back to retransmission.

Fixes: 57f230a ("xen/netfront: raise max number of slots in xennet_get_responses()")
Cc: [email protected]
Signed-off-by: Juergen Gross <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
(cherry picked from commit ad4f15d)

Orabug: 28671425

Signed-off-by: Liam Merwick <[email protected]>
Reviewed-by: Darren Kenny <[email protected]>
Tested-by: Boris Ostrovsky <[email protected]>

Author: jgross1

drivers/net/xen-netfront.c

@@ -1144,7 +1144,11 @@ static RING_IDX xennet_fill_frags(struct netfront_queue *queue,
 			BUG_ON(pull_to <= skb_headlen(skb));
 			__pskb_pull_tail(skb, pull_to - skb_headlen(skb));
 		}
-		BUG_ON(skb_shinfo(skb)->nr_frags >= MAX_SKB_FRAGS);
+		if (unlikely(skb_shinfo(skb)->nr_frags >= MAX_SKB_FRAGS)) {
+			queue->rx.rsp_cons = ++cons;
+			kfree_skb(nskb);
+			return ~0U;
+		}
 
 		xennet_page_ref_inc(queue, page, ref);
 
@@ -1318,6 +1322,8 @@ static int xennet_poll(struct napi_struct *napi, int budget)
 		xennet_page_ref_inc(queue, page, ref);
 
 		i = xennet_fill_frags(queue, skb, &tmpq);
+		if (unlikely(i == ~0U))
+			goto err;
 
 		if (rx->flags & XEN_NETRXF_csum_blank)
 			skb->ip_summed = CHECKSUM_PARTIAL;